{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import os\nimport cv2\nimport random\nimport warnings\nimport numpy as np\nimport pandas as pd\nimport seaborn as sns\nimport matplotlib.pyplot as plt\nfrom sklearn.metrics import confusion_matrix, cohen_kappa_score\nfrom keras.models import Model\nfrom keras import optimizers, applications\nfrom keras.preprocessing.image import ImageDataGenerator\nfrom keras.callbacks import EarlyStopping, ReduceLROnPlateau\nfrom keras.layers import Dense, Dropout, GlobalAveragePooling2D, Input\n\n# Set seeds to make the experiment more reproducible.\nfrom tensorflow import set_random_seed\ndef seed_everything(seed=0):\n    random.seed(seed)\n    os.environ['PYTHONHASHSEED'] = str(seed)\n    np.random.seed(seed)\n    set_random_seed(0)\nseed_everything()\n\n%matplotlib inline\nsns.set(style=\"whitegrid\")\nwarnings.filterwarnings(\"ignore\")","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","_kg_hide-input":true,"_kg_hide-output":true,"execution":{"iopub.status.busy":"2022-05-14T06:31:18.166201Z","iopub.execute_input":"2022-05-14T06:31:18.166582Z","iopub.status.idle":"2022-05-14T06:31:20.638705Z","shell.execute_reply.started":"2022-05-14T06:31:18.166522Z","shell.execute_reply":"2022-05-14T06:31:20.637865Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Load data","metadata":{"_kg_hide-output":true}},{"cell_type":"code","source":"train = pd.read_csv('../input/aptos2019-blindness-detection/train.csv')\ntest = pd.read_csv('../input/aptos2019-blindness-detection/test.csv')","metadata":{"_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_kg_hide-input":false,"execution":{"iopub.status.busy":"2022-05-14T06:31:42.010368Z","iopub.execute_input":"2022-05-14T06:31:42.010693Z","iopub.status.idle":"2022-05-14T06:31:42.035589Z","shell.execute_reply.started":"2022-05-14T06:31:42.010642Z","shell.execute_reply":"2022-05-14T06:31:42.034888Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# EDA\n\n## Data overview","metadata":{}},{"cell_type":"code","source":"print('Number of train samples: ', train.shape[0])\nprint('Number of test samples: ', test.shape[0])\ndisplay(train.head())","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-05-14T06:31:44.136007Z","iopub.execute_input":"2022-05-14T06:31:44.136331Z","iopub.status.idle":"2022-05-14T06:31:44.163362Z","shell.execute_reply.started":"2022-05-14T06:31:44.136272Z","shell.execute_reply":"2022-05-14T06:31:44.162422Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Label class distribution\n\nAs we can see we have an unbalanced database, we have two times more class 0 than 2, and classes 1, 2 and 4 each have less than half of the class 2 data.","metadata":{}},{"cell_type":"code","source":"f, ax = plt.subplots(figsize=(14, 8.7))\nax = sns.countplot(x=\"diagnosis\", data=train, palette=\"GnBu_d\")\nsns.despine()\nplt.show()","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-05-14T06:31:46.439854Z","iopub.execute_input":"2022-05-14T06:31:46.440192Z","iopub.status.idle":"2022-05-14T06:31:46.710345Z","shell.execute_reply.started":"2022-05-14T06:31:46.440138Z","shell.execute_reply":"2022-05-14T06:31:46.709411Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"##### Legend\n- 0 - No DR\n- 1 - Mild\n- 2 - Moderate\n- 3 - Severe\n- 4 - Proliferative DR ","metadata":{}},{"cell_type":"markdown","source":"### Now let's see some of the images\n\nThe images have different sizes, they may need resizing or some padding.","metadata":{}},{"cell_type":"code","source":"sns.set_style(\"white\")\ncount = 1\nplt.figure(figsize=[20, 20])\nfor img_name in train['id_code'][:15]:\n    img = cv2.imread(\"../input/aptos2019-blindness-detection/train_images/%s.png\" % img_name)[...,[2, 1, 0]]\n    plt.subplot(5, 5, count)\n    plt.imshow(img)\n    plt.title(\"Image %s\" % count)\n    count += 1\n    \nplt.show()","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-05-14T06:31:51.408192Z","iopub.execute_input":"2022-05-14T06:31:51.408522Z","iopub.status.idle":"2022-05-14T06:31:59.963613Z","shell.execute_reply.started":"2022-05-14T06:31:51.40847Z","shell.execute_reply":"2022-05-14T06:31:59.95639Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Model parameters","metadata":{}},{"cell_type":"code","source":"# Model parameters\nBATCH_SIZE = 8\nEPOCHS = 20\nWARMUP_EPOCHS = 2\nLEARNING_RATE = 1e-4\nWARMUP_LEARNING_RATE = 1e-3\nHEIGHT = 512\nWIDTH = 512\nCANAL = 3\nN_CLASSES = train['diagnosis'].nunique()\nES_PATIENCE = 5\nRLROP_PATIENCE = 3\nDECAY_DROP = 0.5","metadata":{"execution":{"iopub.status.busy":"2022-05-14T07:15:48.687548Z","iopub.execute_input":"2022-05-14T07:15:48.688138Z","iopub.status.idle":"2022-05-14T07:15:48.696046Z","shell.execute_reply.started":"2022-05-14T07:15:48.688018Z","shell.execute_reply":"2022-05-14T07:15:48.695029Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Preprocecss data\ntrain[\"id_code\"] = train[\"id_code\"].apply(lambda x: x + \".png\")\ntest[\"id_code\"] = test[\"id_code\"].apply(lambda x: x + \".png\")\ntrain['diagnosis'] = train['diagnosis'].astype('str')\ntrain.head()","metadata":{"execution":{"iopub.status.busy":"2022-05-14T06:32:15.687828Z","iopub.execute_input":"2022-05-14T06:32:15.68816Z","iopub.status.idle":"2022-05-14T06:32:15.712098Z","shell.execute_reply.started":"2022-05-14T06:32:15.688101Z","shell.execute_reply":"2022-05-14T06:32:15.711364Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_df = train.groupby(['diagnosis']).apply(lambda x: x.sample(800, replace = True)\n                                                      ).reset_index(drop = True)\ntrain_df[['diagnosis']]","metadata":{"execution":{"iopub.status.busy":"2022-05-14T07:15:13.735411Z","iopub.execute_input":"2022-05-14T07:15:13.73574Z","iopub.status.idle":"2022-05-14T07:15:13.787189Z","shell.execute_reply.started":"2022-05-14T07:15:13.735685Z","shell.execute_reply":"2022-05-14T07:15:13.786307Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"f, ax = plt.subplots(figsize=(14, 8.7))\nax = sns.countplot(x=\"diagnosis\", data=train_df, palette=\"GnBu_d\")\nsns.despine()\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-05-14T07:15:18.413322Z","iopub.execute_input":"2022-05-14T07:15:18.414275Z","iopub.status.idle":"2022-05-14T07:15:18.713446Z","shell.execute_reply.started":"2022-05-14T07:15:18.413926Z","shell.execute_reply":"2022-05-14T07:15:18.712106Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Data generator","metadata":{}},{"cell_type":"code","source":"train_datagen=ImageDataGenerator(rescale=1./255, \n                                 validation_split=0.2,\n                                 horizontal_flip=True)\n\ntrain_generator=train_datagen.flow_from_dataframe(\n    dataframe=train_df,\n    directory=\"../input/aptos2019-blindness-detection/train_images/\",\n    x_col=\"id_code\",\n    y_col=\"diagnosis\",\n    batch_size=BATCH_SIZE,\n    class_mode=\"categorical\",\n    target_size=(HEIGHT, WIDTH),\n    subset='training')\n\nvalid_generator=train_datagen.flow_from_dataframe(\n    dataframe=train_df,\n    directory=\"../input/aptos2019-blindness-detection/train_images/\",\n    x_col=\"id_code\",\n    y_col=\"diagnosis\",\n    batch_size=BATCH_SIZE,\n    class_mode=\"categorical\",    \n    target_size=(HEIGHT, WIDTH),\n    subset='validation')\n\ntest_datagen = ImageDataGenerator(rescale=1./255)\n\ntest_generator = test_datagen.flow_from_dataframe(  \n        dataframe=test,\n        directory = \"../input/aptos2019-blindness-detection/test_images/\",\n        x_col=\"id_code\",\n        target_size=(HEIGHT, WIDTH),\n        batch_size=1,\n        shuffle=False,\n        class_mode=None)","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-05-14T07:15:54.730982Z","iopub.execute_input":"2022-05-14T07:15:54.731349Z","iopub.status.idle":"2022-05-14T07:15:59.98894Z","shell.execute_reply.started":"2022-05-14T07:15:54.731293Z","shell.execute_reply":"2022-05-14T07:15:59.988016Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Model","metadata":{}},{"cell_type":"code","source":"def create_model(input_shape, n_out):\n    input_tensor = Input(shape=input_shape)\n    base_model = applications.ResNet50(weights=None, \n                                       include_top=False,\n                                       input_tensor=input_tensor)\n    base_model.load_weights('../input/resnet50/resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5')\n\n    x = GlobalAveragePooling2D()(base_model.output)\n    x = Dropout(0.5)(x)\n    x = Dense(2048, activation='relu')(x)\n    x = Dropout(0.5)(x)\n    final_output = Dense(n_out, activation='softmax', name='final_output')(x)\n    model = Model(input_tensor, final_output)\n    \n    return model","metadata":{"execution":{"iopub.status.busy":"2022-05-14T07:16:19.781139Z","iopub.execute_input":"2022-05-14T07:16:19.781591Z","iopub.status.idle":"2022-05-14T07:16:19.796849Z","shell.execute_reply.started":"2022-05-14T07:16:19.781528Z","shell.execute_reply":"2022-05-14T07:16:19.793989Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"model = create_model(input_shape=(HEIGHT, WIDTH, CANAL), n_out=N_CLASSES)\n\nfor layer in model.layers:\n    layer.trainable = False\n\nfor i in range(-5, 0):\n    model.layers[i].trainable = True\n\nmetric_list = [\"accuracy\"]\noptimizer = optimizers.Adam(lr=WARMUP_LEARNING_RATE)\nmodel.compile(optimizer=optimizer, loss=\"categorical_crossentropy\",  metrics=metric_list)\nmodel.summary()","metadata":{"_kg_hide-output":true,"execution":{"iopub.status.busy":"2022-05-14T07:16:21.988039Z","iopub.execute_input":"2022-05-14T07:16:21.988383Z","iopub.status.idle":"2022-05-14T07:16:36.624111Z","shell.execute_reply.started":"2022-05-14T07:16:21.988323Z","shell.execute_reply":"2022-05-14T07:16:36.623268Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Train top layers","metadata":{}},{"cell_type":"code","source":"STEP_SIZE_TRAIN = train_generator.n//train_generator.batch_size\nSTEP_SIZE_VALID = valid_generator.n//valid_generator.batch_size\n\nhistory_warmup = model.fit_generator(generator=train_generator,\n                                     steps_per_epoch=STEP_SIZE_TRAIN,\n                                     validation_data=valid_generator,\n                                     validation_steps=STEP_SIZE_VALID,\n                                     epochs=WARMUP_EPOCHS,\n                                     verbose=1).history","metadata":{"_kg_hide-output":true,"execution":{"iopub.status.busy":"2022-05-10T22:03:52.042449Z","iopub.execute_input":"2022-05-10T22:03:52.042718Z","iopub.status.idle":"2022-05-10T22:03:52.380311Z","shell.execute_reply.started":"2022-05-10T22:03:52.042674Z","shell.execute_reply":"2022-05-10T22:03:52.379208Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Fine-tune the complete model","metadata":{}},{"cell_type":"code","source":"for layer in model.layers:\n    layer.trainable = True\n\nes = EarlyStopping(monitor='val_loss', mode='min', patience=ES_PATIENCE, restore_best_weights=True, verbose=1)\nrlrop = ReduceLROnPlateau(monitor='val_loss', mode='min', patience=RLROP_PATIENCE, factor=DECAY_DROP, min_lr=1e-6, verbose=1)\n\ncallback_list = [es, rlrop]\noptimizer = optimizers.Adam(lr=LEARNING_RATE)\nmodel.compile(optimizer=optimizer, loss=\"binary_crossentropy\",  metrics=metric_list)\nmodel.summary()","metadata":{"_kg_hide-output":true,"execution":{"iopub.status.busy":"2022-05-14T07:16:49.650743Z","iopub.execute_input":"2022-05-14T07:16:49.651396Z","iopub.status.idle":"2022-05-14T07:16:49.768152Z","shell.execute_reply.started":"2022-05-14T07:16:49.651111Z","shell.execute_reply":"2022-05-14T07:16:49.767233Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"STEP_SIZE_TRAIN = train_generator.n//train_generator.batch_size\nSTEP_SIZE_VALID = valid_generator.n//valid_generator.batch_size\nhistory_finetunning = model.fit_generator(generator=train_generator,\n                                          steps_per_epoch=STEP_SIZE_TRAIN,\n                                          validation_data=valid_generator,\n                                          validation_steps=STEP_SIZE_VALID,\n                                          epochs=EPOCHS,\n                                          callbacks=callback_list,\n                                          verbose=1).history","metadata":{"_kg_hide-output":true,"execution":{"iopub.status.busy":"2022-05-14T07:17:03.96473Z","iopub.execute_input":"2022-05-14T07:17:03.965071Z","iopub.status.idle":"2022-05-14T07:31:46.591842Z","shell.execute_reply.started":"2022-05-14T07:17:03.965012Z","shell.execute_reply":"2022-05-14T07:31:46.588801Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Model loss graph ","metadata":{}},{"cell_type":"code","source":"history = {'loss': history_warmup['loss'] + history_finetunning['loss'], \n           'val_loss': history_warmup['val_loss'] + history_finetunning['val_loss'], \n           'acc': history_warmup['acc'] + history_finetunning['acc'], \n           'val_acc': history_warmup['val_acc'] + history_finetunning['val_acc']}\n\nsns.set_style(\"whitegrid\")\nfig, (ax1, ax2) = plt.subplots(2, 1, sharex='col', figsize=(20, 14))\n\nax1.plot(history['loss'], label='Train loss')\nax1.plot(history['val_loss'], label='Validation loss')\nax1.legend(loc='best')\nax1.set_title('Loss')\n\nax2.plot(history['acc'], label='Train Accuracy')\nax2.plot(history['val_acc'], label='Validation accuracy')\nax2.legend(loc='best')\nax2.set_title('Accuracy')\n\nplt.xlabel('Epochs')\nsns.despine()\nplt.show()","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-05-10T21:56:56.372252Z","iopub.status.idle":"2022-05-10T21:56:56.372976Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Model Evaluation","metadata":{}},{"cell_type":"code","source":"complete_datagen = ImageDataGenerator(rescale=1./255)\ncomplete_generator = complete_datagen.flow_from_dataframe(  \n        dataframe=train,\n        directory = \"../input/aptos2019-blindness-detection/train_images/\",\n        x_col=\"id_code\",\n        target_size=(HEIGHT, WIDTH),\n        batch_size=1,\n        shuffle=False,\n        class_mode=None)\n\nSTEP_SIZE_COMPLETE = complete_generator.n//complete_generator.batch_size\ntrain_preds = model.predict_generator(complete_generator, steps=STEP_SIZE_COMPLETE)\ntrain_preds = [np.argmax(pred) for pred in train_preds]","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-05-10T21:56:56.374243Z","iopub.status.idle":"2022-05-10T21:56:56.374825Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Confusion Matrix","metadata":{}},{"cell_type":"code","source":"labels = ['0 - No DR', '1 - Mild', '2 - Moderate', '3 - Severe', '4 - Proliferative DR']\ncnf_matrix = confusion_matrix(train['diagnosis'].astype('int'), train_preds)\ncnf_matrix_norm = cnf_matrix.astype('float') / cnf_matrix.sum(axis=1)[:, np.newaxis]\ndf_cm = pd.DataFrame(cnf_matrix_norm, index=labels, columns=labels)\nplt.figure(figsize=(16, 7))\nsns.heatmap(df_cm, annot=True, fmt='.2f', cmap=\"Blues\")\nplt.show()","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-05-10T21:56:56.376031Z","iopub.status.idle":"2022-05-10T21:56:56.376632Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Quadratic Weighted Kappa","metadata":{}},{"cell_type":"code","source":"print(\"Train Cohen Kappa score: %.3f\" % cohen_kappa_score(train_preds, train['diagnosis'].astype('int'), weights='quadratic'))","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-05-10T21:56:56.377701Z","iopub.status.idle":"2022-05-10T21:56:56.378307Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Apply model to test set and output predictions","metadata":{}},{"cell_type":"code","source":"test_generator.reset()\nSTEP_SIZE_TEST = test_generator.n//test_generator.batch_size\npreds = model.predict_generator(test_generator, steps=STEP_SIZE_TEST)\npredictions = [np.argmax(pred) for pred in preds]","metadata":{"execution":{"iopub.status.busy":"2022-05-10T21:56:56.379507Z","iopub.status.idle":"2022-05-10T21:56:56.380087Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"filenames = test_generator.filenames\nresults = pd.DataFrame({'id_code':filenames, 'diagnosis':predictions})\nresults['id_code'] = results['id_code'].map(lambda x: str(x)[:-4])\nresults.to_csv('submission.csv',index=False)\nresults.head(10)","metadata":{"execution":{"iopub.status.busy":"2022-05-10T21:56:56.381111Z","iopub.status.idle":"2022-05-10T21:56:56.381744Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Predictions class distribution","metadata":{}},{"cell_type":"code","source":"f, ax = plt.subplots(figsize=(14, 8.7))\nax = sns.countplot(x=\"diagnosis\", data=results, palette=\"GnBu_d\")\nsns.despine()\nplt.show()","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2022-05-10T21:56:56.383011Z","iopub.status.idle":"2022-05-10T21:56:56.383669Z"},"trusted":true},"execution_count":null,"outputs":[]}]}